The city of Portland is proposing a 1% tax on residents who make more than $125,000 ($250,000 for married filers). This article says that portion of the tax would raise “about $135 million a year.”
I see these kinds of estimates all the time for basically any new tax.
Do these estimates generally take into account that, for example, the implementation of such a tax might cause people who earn in excess of the threshold to move out of Portland? Or are they typically just based on, say, previous year’s tax filings?
Maybe there’s not a general answer.
You’re describing the difference between static and dynamic estimation. Static is easier because it has fewer variables and assumptions, but it risks overlooking the effects you describe. The debate between the two approaches often plays out as to Congressional Budget Office estimates of the effect of new legislation, with the parties generally arguing on a case-by-case basis for whichever approach benefits their position. 
There is a link in the article to a report by EconNorthwest. It estimates between 247 and 274 million as revenue for the proposed taxes. It claims to have adjustments for income growth and avoidance.
Indeed it does - thanks for pointing that out.